Cloudron Forum

I use BuyVM and Hetzner. I've used Digital Ocean in the past.

I use BuyVM for cheap storage that is not all that slow. It's hybrid storage, so it's about $5/TB, and it gets a speed of about 300 MB/sec (only slightly difficult thing is that you have to learn to auto-mount the storage, but this isn't terribly difficult). This is definitely enough speed for a small to medium email server. They are slightly less reliable than the other options (about once every 2-3 months, I have the box go down for an hour or so, and have to restart the server), but they have better support than the rest (you can actually talk to the person who runs the business), so it balances out.

I use Hetzner for NextCloud, because it demands a much higher storage speed in order to feel really 'snappy'. Their SSDs reach about 1600 MB/sec, so it's quite a bit faster, but also more expensive (basically $50/TB). Also, for some reason, I found that even after asking them to enable email, that their servers throttle POP and IMAP responses a lot. I don't suggest running email on them unless you are doing webmail. Their VPSs are very reliable though. Have not tried their full servers.

Digital Ocean is just pretty standard. Slightly expensive and very reliable, so their boxes do the trick if you don't need space.

Also... I forgot to mention this... but thank you Cloudron Team for putting in a 'Disk Speed' option into the Cloudron System UI. It's the best!

@RazielKanos

Yeah, Cloudron could work for this. Cloudron is pretty flexible in terms of how it does emails. Just make sure you use a relay like Amazon SES or Postmark for a bit better deliverability.

For something like this, I'd do it a little differently than @luckow would do it, but pretty similar.

1. Create accounts for users on Cloudron.
2. Create email addresses for each of them (e.g. firstname.lastname@yourdomain.com)
3. Create shared email addresses (e.g. sales@yourdomain.com) and assign permissions to it for the particular users.
4. Have the users log into both their user email addresses (firstname.lastname@yourdomain.com) and their shared email addresses (firstname.lastname@yourdomain.com) using their user account's password. Yep, you do not need a password specifically for your shared email account
5. For Email Clients, you can set up any IMAP client, such as Outlook or Thunderbird. You can also set up SOGO/Roundube/SnappyMail as a webmail client. It's not going to be a smooth as Gmail, but it an be pretty good. (More details on these clients below).
6. To give colleagues access to a mailbox, you can assign them access temporarily in Cloudron, and then revoke it when finished. They can then log in via the webmail client you chose using (your.email@yourdomain.com) and their own personal email account password (not yours).
7. I'll also note that it's possible to set up SOGO and Snappy Mail a bit like Outlook and Thunderbird to access multiple email accounts at the same time. However, as said, SOGO is not quite as smooth as Gmail.

• I hadn't tried Freescount until just a moment ago... it seems like it's great, and it could be a good alternative to shared email inboxes. Personally, I've seen cases where people sort of do a similar thing with synchronizing IMAP Tags in Thunderbird, but this definitely seems a lot more streamlined (in particular, for service cases). It also seems like it'd be a good way to manage permissions for personal inboxes (like I Was talking about in Step 6), but you could have the users do that, rather than have to do that as the administrator. I do wish that they had subfolders though... so I'll probably be sticking with Thunderbird .

• In terms of SOGO v Roundcube v Snappy Mail, in my personal opinion it goes like this. Most of the features on these email clients are the same. SOGO has the most features, followed by Roundcube, and then Snappy Mail. Three features which might or might not be important to you are Multiple Accounts Displayed in a Single Webmail GUI, Vacation Replies, and CARDDAV syncing. SOGO has all three. Roundcube just has vacation replies (but has many other features). Snappy mail has Multiple Accounts, CARDDAV syncing, and Vacation Replies (although the last one is a kinda, as it is not non-technical user friendly to be honest). In terms of how 'quick' the webmail clients feel, it's the opposite. Snappy mail is the fastest, then Roundcube, then Sogo. So it's all a bit of a tradeoff there.

• One final thing to note is that running your own email server like this does take a little bit more work than having someone else hosted. But it's also a great deal cheaper, especially when you need more accounts. So I personally find it worth it, but your mileage may vary.

Hey there,

I'm not a paying customer (yet) so don't feel any obligation or pressure to do something like this... I'm not sure if your customers actually need something like this actually. Although I'd find this useful myself.

Currently, for backup retention, there is a dropdown of different types of retention schedules. These retention schedules are quite varied, and the ones towards the bottom are more geared towards the rsync backup option, I believe.

I like the options further down in the list (like 2 daily, 4 weekly) because they provide options if someone messes something up and you don't notice it for a while. However, currently I've been sticking with the tarball option because it's just a lot faster for backing up things like NextCloud. The downside though, is that tarballs are a lot larger, and running a 2 daily 4 weekly tarball backup uses up a lot of space.

I was wondering if it was at all possible for Cloudron to offer a few more retention policy options... things like '2 daily 2 weekly' (although I'm not necessarily sure this is the particular one you guys want). Whether that's by adding more options in the dropdown, or allowing us to set a specific number of daily, weekly, and monthly backups retained - I think either would be pretty awesome.

In any case, let me know what you think, and thank you for all of the work that you do. Your software is really great, and thank you for all of the hard work you guys do making this software which makes installing and backing up these apps way way easier.

Hey all, the most recent Nextcloud update has been out for a bit, but I wanted to give those who haven't updated yet a heads up.

If you happen to have an installation with a large amount of Users (regardless of whether or not they use it or not), expect the container to take a while to go from 'Starting' to 'Running' when you boot it for the first time after updating.

This is because Nextcloud is attempting to generate a bunch of avatars for some reason, as part of the new update. To give you an idea, it was probably taking 5-10 seconds per user for me. I was confused at first, and so I restored from backups a ton of times. You don't need to do that. Instead, just give it the time to run through it's process and it will go to 'Running' just fine.

Anyways, here's where I got the information from, after experiencing it myself: Thread on Nextcloud Forums

Also, as a note, if anybody is having trouble with add-ons after an update, sometimes, the problem can be resolved by uninstalling the disabled add-on (after the update), and then re-installing it. For some reason, that seems to get it working again (whereas just re-enabling it will not). Every time I've done it, it keeps the configuration settings for the add-ons just fine, although it may be good to manually take a screenshot of those addon settings before doing the removal, just in case.

Just wanted to give everyone an update.

Last week was pretty crazy for me, so I wasn't able to do much testing.

I will probably resume doing some more tests this week. In any case, restarting dovecot every hour does seem to be a suitable mitigation for anybody who has this issue (at least pending until a solution is found).

@jayonrails Hey there,

In general, the easiest thing to do is to do a remote backup by deafult, as girish said. If you aren't very familiar with Linux servers, it is probably best to stick with this. In terms of integrity... I don't think rsync provides any guarantees there. You might have to write something on your own. For prepping against an unlikely hack or malware, your best bet is to stick it on a backup storage provider that does snapshotting, like rsync.net (unlikely to happen though, as Cloudron takes security pretty seriously).

However, if you need a local backup in addition to a remote backup, the easiest thing to probably do is to use Cloudron Backup in rsync mode to rsync to disk, and then to use restic to backup that directory to a backup storage service. Note that it is very important that the disk you rsync to is NOT the same one Cloudron is running on. If you are using a VPS provider, put it on one of their 'block storage' disks, or something like that.

Having the local backup -> remote backup workflow helps in the sense that it's a lot quicker to do a backup locally than it is remotely, and restic packs those backups remotely really efficiently (and reasonably quickly too).

This does require some knowledge on crontabs and scripting though. There are some forum posts on the general process. Here's one, though they backup to one drive:

https://forum.cloudron.io/topic/6928/tutorial-remote-backup-of-local-cloudron-backup-snapshots-with-restic-rclone?_=1676593563170

What I've done personally is used a CLI tool called autorestic to help manage the backups. Then I've written some short shells scripts that can do a restore, list backups, and mount backups as a drive (a feature of restic). This makes it easy for me to go through my backup's files, even though they are remote and encrypted, which is nice.

You do take on a lot more responsibility doing it this way though, as you are responsible for the scripts - you are on your own to make sure things work. So best of luck to you, if you choose to do things this way :).

Snapshots are just a secondary way to preserve your system. Let's say someone breached your VPS, trashed it, and managed to grab the key to your backups, and then went an trashed your backups (This would take a very determined actor, but let's suppose). You could role back to a previous snapshot most of the time (however snapshots while a VPS is running live can boot up to a broken system sometimes).

As a note, some backup providers also do snapshotting, which would allow you to recover your backups even if someone went in and thrashed them (rsync.net does this for instance).

You'll need to decide how much protection is important for you. Just keep in mind that snapshots are not a backup, just like RAID isn't a backup either :).

@jdaviescoates Woops! Fixed

@nebulon @jdaviescoates @chetbaker

Thanks! I feel pretty dumb here, as the reason I had this problem was that I just didn't start a meeting. I assumed (incorrectly) that it would block the main page with a log in screen, and not just the 'I am host' button once you actually start a meeting. Thank you for the answer :).

Hey Everyone,

I was just wondering what VPS's you all use for hosting your Cloudrons? I figured it might be helpful for those who are looking to get started in this.

Currently I use BuyVM for my VPS, and use their very cheap Block Storage ($5/TB/month!) for a ton of cached hard drive space. It works really well actually.

For backups, I am currently trying that with a Hetzner Storage Box currently using a NexusBytes box for backup (much closer to the VPS than Hetzner was, which means speeds are about 20x faster). Would have used BuyVM for this as well except that I like to host the VPS and Backup on different services.

What does everyone else use?

@girish This makes a ton of sense, although I'm not quite sure if I'm doing it right. Under what you linked to, the section should be under the Resources tab, although I think it's moved to the Storage tab (no problems with that, I just wanted to make sure I'm handling it correctly).

The part that's a little confusing is how 'App Data' and Volumes tie together in storage (now it seems this may be resolved as you said, since your working on Volume updating at some point in the future).

Anyways, App Data says to mount an external EXT4 disk, which I assume is done with the 'Mounts' interface right below it. However, the mount says that you are mounting 'Volumes'.

The documentation says 'App Data' is backed up. However, the documentation also says that 'Volumes' are not backed up. So that does leave me a little perplexed as to whether the data will be backed up or not. I think according to your comment, that it will be backed up (because you did say it would no matter where it is on the server). But I just thought I'd check

In any case, thanks for your help, and hope you have an awesome rest of your week.

@humptydumpty Hey, so I have some Feedback on the Hetzner Storage Box. If you use it in Europe, it'll probably be all right, as you are a lot closer to the box.

However, for US Backup Storage of any sizable amount, it's going to be very slow. I was looking at around 60 days or so to upload 300GB (probably mostly because of the distance). So I'm looking for something a bit closer.

My ping response time for Hetzner was 146 ms, which isn't terrible... but I can get something around 7 ms with a VPS closer to home, which I think will work a lot better for the backup. We will see though.

@d19dotca I can't believe I missed that. Literally read right past it. Thank you!

@marcusquinn said in Email deliverability to Microsoft email servers:

Anyone know if the sent mail is still stored on the IMAP server before it's sent to the relay to deliver?

This depends on your email client settings actually. Basically, most email clients save a copy of the email you are sending in the 'Sent' folder, but they can be configured not to. The saved copy is uploaded to your actual 'mailbox' via IMAP - but this is not equivalent to sending an email, this is just saving a copy that doesn't leave your mailbox. IMAP is basically what allows you to organize and upload and download to that mailbox - and it doesn't have anything to do with actually sending of the email.

Actual outgoing email goes through a different protocol called SMTP - and it doesn't touch your actual 'mailbox' at all except for when email is delivered to you. SMTP is used by your email client to send your email to your own email server, and then SMTP is used again from your server to the receiver's email server (and sometimes multiple chains if it's configured that way). So it's basically just a protocol for sending email.

Anyways, best of luck with all of this . Sorry it's being so painful to get it working.

@BrutalBirdie Hey Brutal, actually if it's cool, let me do a few tests first. I think I might have isolated a particular issue... and it might be unique to thunderbird... if it's just a thunderbird issue, it might not be related to the email server at all.

I'll let you guys know later today... unfortunately, today is likely to be crazy busy... just did a really large email migration, and those are kinda painful sometimes :).

My use case is a little different than yours, so I don't have a tutorial, but Directus shouldn't be terribly hard to integrate with a JAM website.

I'm currently using it as a source for a static website using Elder.js. Elder.js isn't a JAM - it's a static site generator - but it works kinda similarly. Directus works wonderfully for it.

In most of the JAM stacks, there is a point for fetching data, and at that point (assuming your using Node.js here), you just use axios to call the endpoints that Directus exposes. When you want to create user accounts or whatnot, you make an API call to Directus.

I don't know how well Directus would stand up to a busy traffic website performance wise, but I'm pretty sure it would do a great job for a low-to-moderate traffic website. It might be a good thing to ask them on their Discord, they are very friendly.

A good way to picture Directus is that it's a nice frontend to an SQL database. So a lot of the things you'd care about with a regular database will matter with Directus. It's not going to automatically manage the accounts for you - but it's not hard to create a collection called 'Website Users' and just add a user account when someone registers on your site, and then verify it when they log in, managing the cookie or token appropriately.

This of course means it isn't 'low-code', but it isn't 'that's a ton of code' either. (As a note, if your going for posts and logins and the likes, I'd stick with JAM, and stay away from static site generators - static site generators are really built for sites that don't have live content, not that you can't fenagle your way around it, but it's not what they are built for).

If your looking for a bit more automation then that for logins... there's a few other frameworks that have built in libraries for user authentication, but the only one I remember really standing out as being 'easy' was Meteor (using their old UI framework called 'Blaze'). But Meteor is pretty different from Directus, and it's very opinionated, and usually requires a heavier server. Might not be what you are looking for.

One thing I will warn you on Directus... sometimes, major version upgrades (think 8.?.? -> 9.0.0) don't work quite as planned (at least in my experience). So if you see they are going to put out a new major version soon, it might be worth holding out until they do. Just my personal experience though.

Best of luck with your project!

Hey all,

I was hoping some of you might have a bit of guidance on some email issues I've been having. If not, that's cool too though :).

I'm running about 300G of email on a Cloudron System. Currently, right now, I'm running into a case where sporadically, some IMAP and POP folder reads will time out. These aren't huge folders either - some only have 2-3 files. SMTP is not affected.

There's some extraneous situations which might be related to this. Two slightly problematic things happened over the past few weeks. (1) I ran an external backup disk out of storage - it isn't tied to the disk the mail is on in any way. (2) I had to Hard-stop the server on the VPS provider side of things because of a configuration issue.

I'm not entirely sure which of the two instances caused the issue, but it has been occurring ever since. I'm wondering if some files got corrupted or something.

One other thing I've noticed while doing this is that the mail docker instance is taking up a lot more memory. I've raised it from 500K to 1G to 2G, and sometimes it still even fills up that. I'm not sure if that's related or not - but the issue does occur whether it is full or not - I've had cases where the issue occurs when only using about 1G memory out of 2G allocated to it. It does tend to happen a bit more when it's full though.

When I've run a Cloudron Email Server in the past, it's not really had this issue of memory filling up, but they've also been for far less email (10G or so). Then again, those have all been sending email directly, not through SES.

Restarting the mail process fixes the issue until it happens again. However, it's very difficult to predict when the issue will happen again.

I've gone into the mail image inside and edited the dovecot config and turned on mail_deubg and auth_debug - I'll post some information from the logs when I get some information from this. EDIT: Not seen anything significant in logs. Will probably wait for another occurrence and see if I can spot any patterns.

I'm trying not to rely on the Cloudron Support team because I have not bought Premium yet (will probably do in the future, but I'm not quite ready to do so at the moment).

Have any of you seen issues like this before? If so, how have you proceeded towards fixing it?

@neoplex Just as a heads up, I believe Uptime Kuma does a lot of the same stuff. If you're looking for something they currently have on Cloudron right now (there is not an import tool though...)

Another update. Updated the 2007 Outlook Client to Office 365. Haven't had any issues with them so far, but need to wait to be sure. There's a small chance that this whole thing was actually two separate email client issues, one with Thunderbird, and one with Outlook.

Certain versions of Outlook actually have an issue where they don't respond to Dovecot's request to see if they are 'still there' and if an IDLE IMAP connection is closed, they will go offline automatically (https://wiki.dovecot.org/Timeouts and also see point 2 and 3 of the last post in https://social.technet.microsoft.com/Forums/exchange/en-US/b418d42a-ce66-4cfa-adb9-e6e24ec769f4/outlook-2007-error-your-imap-server-closed-the-connection). I'm thinking this might have been the issue. Not seeing this issue on Office 365 clients other users are running.

As for the Thunderbird issue... found one account of someone running into something similar here: https://www.mail-archive.com/dovecot%40dovecot.org/msg82805.html . Still gotta work on that one, and do some measurements to see if we are hitting the process limit when we have the issue. This would be a weird case though as we are only seeing this issue with some Thunderbird clients, despite all running the same version and with similar software on the computers... which again points to a mailbox level issue. Gotta do some more sleuthing.

Hey all, reinstalled and then did a restore to fix things. Had a few trouble with restore, but was able to work them out. Also wrote a quick rsync command to take the emails from lost+found and restore them to their proper directories, so no emails lost :).